Increased Differentiation of Dermal Mast Cells in Mice Lacking the Mpl Gene

Abstract

Thrombopoietin interactions with its receptor, Mpl, play an important role in the regulation of hematopoietic stem/progenitor cell proliferation and differentiation. In this study, we report that the mast cell restricted progenitor cells (MCP) and the mast cell precursors in the bone marrow of wild-type mice express Mpl on their surface. Furthermore, targeted deletion of the Mpl gene in mice decreases the number of MCP while increasing the number of mast cell precursors present in the marrow and spleen. It also increases the number of mast cells present in the dermis, in the peritoneal cavity, and in the gut of the mice. In addition, serosal mast cells from Mpl^null mice have a distinctive differentiation profile similar to that expressed by wild-type dermal mast cells. These results suggest that not only does ligation of thrombopoietin with the Mpl receptor exert an effect at the mast cell restricted progenitor cell level, but also plays an unexpected yet important role in mast cell maturation.

Introduction

Mast cells are stem cell–derived hematopoietic cells that have a dual role of immune effector against parasitic invasion as well as mediator of the allergic response [1–2]. These cells arise in the bone marrow and to a lesser extent in the spleen and reside in highly vascularized extramedullary sites, such as the dermis of the skin and the mucosa of the gut [3–5]. Hematopoietic progenitor cells committed to mast cell lineage were first identified in the fetal blood of a mouse [6]. More recent data have identified a progenitor cell population restricted to the “mast cell” lineage, the MCP, in the marrow of adult mice with the phenotype Lin^negc-Kit^posSca-1^negLy6c^negFcɛRIα^negCD27^negβ7integrin^posT1/ST2^pos [7]. The precise relationship of the MCP in the hematopoietic progenitor cell hierarchy is debated [8]. Some investigators have proposed that the MCP is derived from the common myeloid progenitor cell, the CMP [7], while others argue that in addition these cells may derive from the granulocyte-monocyte restricted progenitor cell, the GMP [9]. Under normal circumstances, MCP do not differentiate in the marrow, as indicated by the fact that the frequency of mast cell precursors (c-Kit^highCD34^pos) in this tissue remains limited (0.02%) throughout adult life [10]. Instead, MCP circulate in the blood to colonize extramedullary sites where they differentiate into tissue-restricted mast cells, each with a specific mast cell protease (MMCP) (for a review on mast cell–specific MMCP, see [11,12]) expression profile [13,14], giving rise to dermal, mucosal, and serosal mast cell populations. The molecular mechanism in normal mice that restricts the mastocytopoietic potential of stem/progenitor cells to extramedullary sites, as well as the factors that guide their differentiation along different lineages is unknown.

There are several similarities between the pathways that control megakaryocytic and mast cell differentiation: Differentiation of both lineages is dependent on the growth factor stem cell factor (SCF) (in synergy with thrombopoietin, TPO [15], and IL-3 [3–5], respectively), and the transcription factors Gata2 and Gata1 [16–18]. In both lineages, the expression of Gata1 is regulated by the first enhancer, the DNase hypersensitive site I, of the gene [16,17], a proposed target for TPO signaling [19,20]. Furthermore, three independent studies have reported that TPO increases the cellular output in BMMC cultures seeded with human CD34^pos cells [21–23] and genetic alterations of TPO signaling are associated with the development of myeloproliferative disorders expressing both megakaryocytic and mast cell abnormalities [24]. This would suggest that TPO, a growth factor produced by the osteoblasts [25] and present in the marrow stem cell niche [26], might be involved in the regulation of mast cell differentiation. Wild-type serosal murine mast cells have been shown in previous work to express the mRNA for Mpl, the receptor for TPO [20], and in vivo mouse treatment with TPO or addition of this growth factor to bone marrow–derived murine mast cell cultures strongly decrease the generation of mature mast cells by inducing apoptosis [27]. To further clarify the effects of TPO on mast cell differentiation, we show that MCP and mastocytic precursors present in the marrow of mice express Mpl on the cell surface. Next, we determined that targeted deletion of the Mpl gene in mice has effects on mast cell differentiation opposite to those described previously with TPO treatment (27): decreasing the number of MCPs present in marrow and spleen, and increasing the number of mast cell precursors in the connective tissue, in the mucosa, and in the peritoneal cavity. In addition, mast cells derived from Mpl^null mice have a protease expression profile similar to that of dermal rather than serosal wild-type mast cells. Finally, we show that, in mice, TPO is expressed by cells from multiple organs including, in addition to liver and marrow, spleen, kidney, and gut. Moderate bleeding, or the presence of the Mpl^null mutation, decreases the level of TPO protein detected in blood while altering the levels of TPO mRNA expressed in most of the other tissues. Altogether, these results confirm that TPO plays an important role in selectively increasing dermal versus mucosal mast cells.

Materials and Methods

Mice

Mpl^null mice [28] were provided by Dr. W. Alexander (Walter and Eliza Hall Institute for Medical Research, Melbourne, Australia) and bred with CD1 females (Charles River Laboratories, Inc., Calco, Italy) at the animal facilities of Istituto Superiore Sanità. Littermates were genotyped by PCR and those found not to carry the mutation used as wild-type controls. In selected experiments, wild-type mice were bled by harvesting 400 μL of blood through the retro orbital plexus with a glass Pasteur pipette. All the experiments were performed with 6- to 10-month-old male littermates, according to protocols approved by the institutional animal care committee.

Histological analysis

The ear, spleen, femur, and stomach were fixed in 10% (v/v) phosphate-buffered formalin (Sigma, St. Louis, MO), paraffin embedded and cut into 2.5–3 μM sections according to standard procedures. Slides of consecutive sections were dewaxed, rehydrated, and stained with regular and acidified Toluidine Blue (Multilab, Surrey, England) and safranin-counterstained Alcian Blue (Bio-Optica, Milano, Italy), as described in [16]. Cell metachromasia is defined by the color acquired by the cytoplasmic granules after acidified toluidine blue staining. The granuli of immature and mature mast cells are blue (metachromatic^neg) and red (metachromatic^pos), respectively [16]. Light microscopy was analyzed using a Leica Light Microscope (Leica Microsystems Ltd., Heidelberg, Germany) equipped with a Coolsnap videocamera for computerized images (RS Photometrics, Tucson, AZ) while transmission electron microscopy was performed using the EM 109 Zeiss (Carl Zeiss Optronics GmbH, Oberkochen, Germany).

Flow cytometry analysis and cell purification

Cells were labeled with PE-CD117 (that recognizes c-Kit) coupled with either fluorescein isothiocyanate (FITC)–CD34, FITC–Annexin V, or FITC–CD45R/B220 as control. Further antibodies used in this study are FITC-Mac3, -CD61, -T1/ST2 (MD Biosciences Inc., St. Paul, MN) and -CD71 and PE-Ly-6G (Gr1), -CD41, -TER119. The expression of FcɛRI was revealed by sequential incubations with the monoclonal mouse anti-DNP-IgE (clone SPE-7; Sigma) and FITC-conjugated rat antimouse IgE [6]. Expression of Mpl was analyzed by three color flow cytometry using a biotinylated Mpl antibody (AMM2; Kirin Pharmaceuticals, Takasaki, Japan) and APC-CD117, coupled either with FITC-FcɛRI (eBiosciences, Inc., San Diego, CA) or FITC-T1/ST2 [29]. In these experiments, aspecific antibody binding was reduced by preincubating the cells with a FcγRII/III blocker (CD16/CD32). Unless otherwise stated, all the antibodies were from PharMingen (PharMingen, San Diego, CA) and were incubated at a concentration of 1 μg/10⁶ cells for 30 min on ice. Cell fluorescence was analyzed either with a Coulter Epix Elite ESP (Beckman Coulter, Inc., Miami, FL) or a FACS ARIA (Becton Dickinson, Franklin Lakes, NJ). Nonspecific fluorescent signals were gated out with appropriate fluorochrome-conjugated isotype controls and dead cells were excluded by propidium iodide staining. Serosal mast cells were purified as described in [16]. Briefly, peritoneal cells were first incubated with FITC-conjugated CD45R/B220 and immunodepleted with a monoclonal mouse anti-fluorescein antibody–coated MACS microbeads (Miltenyi Biotech GmbH, Bergisch Gladbach, Germany), according to the manufacturer's instructions. The B220-negative cell fraction was then incubated with PE-CD117, and CD117^high cells (≈50% of B220^neg cells) isolated with the FACS ARIA (>95% CD117^high after reanalysis).

TPO ELISA

Bone marrow washes were prepared by gently flushing the cavity of two femurs with 400 μL of phosphate buffered saline containing 0.1% Tween as described in [30]. The same buffer was used to prepare cell washes from liver and spleen. Blood was collected from the retro orbital plexus of the animals with a heparin-coated Pasteur pipette and platelet-poor plasma prepared by centrifugation at 5000 rpm (platelet contamination of the samples <9%). The levels of TPO present in plasma and cell washes were measured using a commercial ELISA kit (Quantikine Immunoassay mouseTPO; R&D Systems, Inc., Minneapolis, MN).

RNA isolation and quantitative RT-PCR

Total RNA was prepared with Trizol (Gibco-BRL, Grand Island, NY) and reverse transcribed with random primers using the SuperScript III kit (Invitrogen Life Technologies, Bethesda, MD). Gene expression was quantified using the Taqman PCR kit (PE Applied Biosystems, Foster City, CA) and predeveloped custom-made oligos, whose sequence is available upon request, using either the ABI PRISM 7700 or the 7300 Sequence Detection System (Applied Biosystems) [27,29]. GAPDH was also quantified in each reaction. Results were analyzed by using the SDS program (v1.9; Applied Biosystems) and expressed in arbitrary units, using the amplification of GPDH as calibrator, according to the following algorithm: ΔCt = [CtX–CtGPDH], where Ct is the threshold cycle of the gene analyzed, and are presented as 2^−ΔCt.

Statistical analysis

Statistical analysis was performed by analysis of variance (ANOVA) using Origin 6.1 software for Windows (Microcal Software Inc., Northampton, MA).

Results

Mpl is expressed by mast cell restricted progenitor cells (MCP) and precursor cells present in the mouse bone marrow

By using flow cytometry analysis, the anti-Mpl antibody clearly labels above isotype control levels myeloid progenitors (MP, CD117^pos/T1/ST2^pos) and MCP (CD117^pos/T1/ST2^pos) from wild-type mice. The levels of Mpl expression on the surface of MCP are significantly higher than those expressed by MP (1658 ± 73 vs. 163 ± 73, respectively) (Fig. 1A). The anti-Mpl antibody clearly labels above isotype control also mast cell precursors (CD117^pos/FcɛRI^pos) in the marrow of wild-type mice (Fig. 1B). The levels of expression on Mpl on the surface of these precursors is lower than those expressed by the MCP (868 ± 71) and comparable to those expressed by wild-type megakaryocytes (CD41^pos cells) analyzed as positive controls (958 ± 74) (Fig. 1B). In contrast, Mpl is not detectable on the surface of megakaryocytes and mast cell precursors from the marrow of Mpl^null mice used as negative control (Fig. 1B).

FIG. 1.

The Mpl receptor is expressed on mast cells and mast cell–restricted progenitor cells present in the murine marrow. (A) Fluorescence-activated cell sorter (FACS) analysis with a biotinylated-irrelevant isotype (PE-Cy7-A) (top panels) or biotinylated-AMM2 (PE-Cy7-A anti-Mpl) antibody (bottom panels) of wild-type progenitor cells (c-Kit^pos). Progenitor cells were divided into myeloid restricted (MP, blue) and mast cell–restricted (MCP, red) progenitor cells on the basis of T1/ST2 staining [7]. The gating defining MP and MCP was confirmed by a functional assay based on the ability of prospectively isolated cells to generate myeloid or mast cell precursors, respectively, in culture [29]. The cells were incubated with the antibodies in the presence of a FcγRII/III blocker. The mean (±SD) peak fluorescence intensities (in arbitrary units), observed in three separate experiments, are indicated. **Significantly (P < 0.01) higher than isotype controls. (B) FACS analysis with a biotinylated-irrelevant isotype (PE-Cy7-A) or biotinylated-AMM2 (PE-Cy7-A antiMpl) antibody of mast cells (CD117^posFcɛRI^pos) and megakaryocytes (CD41^pos) from the marrow of wild-type (left) and Mpl^null (right) littermates. The cells from Mpl^null mice represent negative controls. The cells were incubated with the antibodies in the presence of a FcγRII/III blocker. The mean (±SD) peak fluorescence intensities (in arbitrary units), observed in three separate experiments, are indicated. **Significantly (P < 0.01) higher than isotype controls (gray histograms).

In conclusion, in contrast with the restriction of Mpl expression that occurs during lymphoid commitment [31], the expression of Mpl increases during the transition from MP to MCP and decreases when MCP mature into mast cell precursors to levels comparable to those expressed by megakaryocytes.

Targeted deletion of the Mpl gene increases the number of mast cells and of MCPs present in the marrow

The observation that mast cells and their progenitors express Mpl prompted us to analyze whether targeted deletion of this gene would alter mastocytopoiesis in mice. The number and morphology of mast cells and of their progenitors present in bone marrow and spleen of wild-type and Mpl^null mice is compared in Figure 2.

FIG. 2.

The marrow from Mpl^null mice contains few mast cell restricted progenitor cells (MCP) (CD117^pos/T1/ST2^pos) but numerous mast cells (CD117^posFcɛRI^pos). (A) Alcian Blue/Safranine staining of marrow sections from wild-type and Mpl^null mice. The insert in the panel of the right represents a representative Alcian Blue/Safranine positive cell. Magnification: (A and B) 10× and (insert) 40×. (B–C) FACS analysis for CD117/T1/ST2 (B) and CD117/FcɛRI expression (C) of marrow and spleen cells from wild-type and Mpl^null littermates, as indicated. CD117^pos/T1/ST2^pos and CD117^posFcɛRI^pos profiles are defined according to the gates indicated by rectangles. Side (SSC) and forward (FSC) scattogram analyses divide CD117^posFcɛRI^pos cells into nGHP_S, nGHP_L, and GHP_L, corresponding, respectively, to nongranular-small, nongranular-large, and granular-large mast cells, as described in [10]. The mean (±SD) frequency of Alcian Blue/Safranine^pos (A), CD117^posT1/ST2^pos (B), and CD117^posFcɛRI^pos cells observed in at least three separate experiments are presented on the right. Values statistically different (P < 0.01) from wild-type controls are indicated by asterisks. The values for wild-type mice used in these experiments are the same as those published in [27].

As expected on the basis of results with other progenitor cell types [28], the marrow from Mpl^null mice contains lower numbers of MCP (Fig. 2B). At the precursor level, marrow sections from Mpl^null mice do not contain granulated cells detectable by Toluidine Blue staining ([10,32] and data not shown), but contains numerous cells stained golden-brown by Alcian Blue/Safranine, indicating the presence of mast cells with an immature phenotype in these mutants (40 ± 3 vs. 2 ± 1 cells/mm² in Mpl^null and wild-type marrow, respectively) (Fig. 2A).

The mast cell nature of the Alcian Blue/Safranine^pos cells is confirmed by flow cytometry for CD117 and FcɛRI co-expression. The CD117^posFcɛRI^pos cells comprise only 0.2% of the total bone marrow cell population in wild-type mice. As reported previously [10,32], these cells are divided, by side and forward scatter analysis, into nongranular-small (nGHP_S) and granular-large (GHP_L) cells (Fig. 2C). Also the spleen from normal mice contains few (0.5%) CD117^pos FcɛRI^pos cells (Fig. 2C). In contrast, the bone marrow and spleen from Mpl^null mice contain 10 times more (10–19%) CD117^posFcɛRI^pos cells than the corresponding tissues from wild-type littermates (Fig. 2C). The immature nature of the CD117^posFcɛRI^pos cells in the marrow is confirmed by the fact that these cells express high levels of mast cell carbonic peptidase A (MC-CPA) (2^−ΔCt = 6.4 ± 0.2 × 10⁻²) but barely detectable levels of MMCP-6 and MMCP-7.

Targeted deletion of the Mpl gene increases the number of dermal, mucosal, and serosal mast cells and alters the differentiation profile of serosal mast cells

Mpl^null mice contain 60% more of immature (metachromatic^neg after Toluidine Blue) and ≈2.5-fold more of mature (Alcian Blue/Safranine^pos) mast cells in the dermis of the ear than wild-type animals (Fig. 3). On the other hand, the frequency of TUNEL^pos dermal mast cells in the ear from Mpl^null and wild-type littermates is comparable (Fig. 3C). By electron microscopy, dermal Mpl^null mast cells contain more highly electron dense granules than the corresponding wild-type cells (Fig. 4). Increases in the frequency of immature and mature mast cells are also observed in the connective region of the stomach of Mpl^null mice (Fig. 5). Quantitative measurements of the longer diameter of the mast cells indicated that Mpl^null dermal, mucosal, and serosal mast cells are ∼2-fold bigger than the corresponding wild-type cells (Fig. 6).

FIG. 3.

The dermis from Mpl^null mice contains many immature (metachromatic^neg after Toluidine Blue) (A) and mature (Alcian Blue^pos/Safranine^pos) (B) mast cells. Ear sections were also stained using TUNEL, as presented in (C). Photograph of representative sections from the ears of wild-type and Mpl^null mice are presented on the left, while the mean (±SD) frequency of cells observed in sections from three separate animals per experimental point is presented on the right. Values statistically different (P < 0.01) from wild-type controls are indicated by asterisks. The values for wild-type mice used in these experiments are the same as those published in [27]. Magnification: 40×.

FIG. 4.

By using electron microscopy, Mpl^null mast cells present more highly dense heterogenous cytoplasmic granules than wild-type mast cells. The number and electron density of granules present within the cytoplasm of the mast cells were quantified by analyzing with the Methamorph program at least 10 separate wild-type and Mpl^null cells. Values statistically different (P < 0.01) from those observed with wild-type controls are indicated by asterisks. Mast cells and granules are presented at 4400× and 30,000× magnification, respectively.

FIG. 5.

The connective tissue from the stomach of Mpl^null mice contains numerous immature (metachromatic^neg after Toluidine Blue) and mature (Alcian Blue^pos/Safranine^pos) mast cells. The bar graphs on the right show the significant increase (P < 0.01) in the number of mast cells/mm² in the stomach from Mpl^null mice with respect to those found in wild-type mice. The results are presented as mean (±SD) of independent determinations with three different wild-type and Mpl^null mice. Mast cells are indicated by arrows. Magnification: 20× in the panels and 40× in the inserts.

FIG. 6.

Mast cells from Mpl^null mice are bigger than those from wild-type animals. Results are presented as mean (±SD) length of the larger diameter of 10 randomly selected mast cells identified in histological preparation of ears, stomachs, or peritoneal lavages (PL) of three separate wild-type and Mpl^null mice. Values statistically different from those observed in wild-type mice are indicated with an asterisk. The diameter was measured using the Methamorph program.

Mpl^null mice also contain numbers of CD117^highFcɛRI^pos mast cells higher (by 3- to 6-fold) than normal in the peritoneal cavity (Fig. 7). The increase is due both to higher frequency of CD117^highFcɛRI^pos cells (4.3 ± 2.3% vs. 1.0 ± 0.1%, respectively) and to higher numbers of cells present in the peritoneal cavity (≈15 × 10⁶ vs. 8 × 10⁶ cells/mouse, respectively). Serosal Mpl^null mast cells are bigger (Fig. 6), contain more granules, react more strongly with Alcian Blue (Fig. 8), and express higher level of CD117 (36,700 ± 6500 vs. 8700 ± 1100 AFU/cell, respectively, P < 0.05) and FcɛRI (two populations expressing either 600 ± 30 or 1870 ± 760 AFU/cell vs. one population with 120 ± 10 AFU/cell, respectively, P < 0.05) (Fig. 7) than the corresponding wild-type cells (Fig. 7).

FIG. 7.

The Mpl^null mutation increases the number of serosal mast cells. FACS analysis for CD117/FcɛRI, CD117/Annexin-V, or CD117/irrelevant antibodies (negative controls, gray curves, and dotted lines) of cells from the peritoneal lavages of wild-type and Mpl^null littermates, as indicated. The mean (±SD) frequency of c-Kit^pos cells, the average (±SD) fluorescent intensity of the FcɛRI signal and the frequency of Annexin^pos cells are also indicated. Magnification: 40×.

FIG. 8.

The Mpl^nul1 mutation alters the differentiation state of serosal mast cells. Expression profiling by quantitative RT-PCR of CD117^highFcɛRI^pos cells purified from wild-type and Mpl^nul1 littermates. Gates used for purification and reanalysis of sorted cells for purity (>95%) are presented on the top. May-Grunwald (top panels) and Alcian Blue (bottom panels) staining of purified cells are also present as control. Magnification: 40×. Mean (±SD) of three separate experiments performed in triplicate. *(P < 0.05), **(P < 0.01), statistically different from wild-type controls.

By quantitative RT-PCR analyses, wild-type and Mpl^null serosal mast cells are also different. Wild-type serosal CD117^pos cells express high levels of c-Kit (2^−ΔCt = 1.4 ± 0.2 × 10⁻²), high levels of Mitf and Gata2 (2^−ΔCt ≈1.4 × 10⁻² for both), low levels of Gata1 (2^−ΔCt = 3.1 ± 0.3 × 10⁻⁴), high levels of MC-CPA (2^−ΔCt = 9.9 ± 1.0 × 10⁻²) and MMCP-6 (2^−ΔCt = 1.2 ± 0.1), and relatively low levels of MMCP-7 (2^−ΔCt = 2.1 ± 0.2 × 10⁻²) (Fig. 8), a protease mainly produced by dermal mast cells [33]. On the other hand, serosal mast cells from Mpl^null mice express levels of c-Kit, Gata1, Gata2, Mitf, Bcl-2, and MMCPs 0.5–2–log, higher than those expressed by the corresponding wild-type cells (Fig. 8).

Decreases in blood platelet numbers alter the pattern of TPO expression in the extramedullary tissues of mice

It is currently accepted that TPO is mainly produced in the liver [34] and to a less extent in the bone marrow [25]. Recent evidences indicate, however, that TPO is expressed in a variety of other organs [35]. For this reason, we investigated the levels of TPO expressed in organs (stomach, intestine, spleen, and ear) rich in mast cells using quantitative RT-PCR and ELISA. As platelets exert multiple controls on the levels of TPO levels expressed in vivo [25,34–36], the amount of TPO expressed by tissues of mice in which the number of platelets has been reduced either by bleeding (as a model of acute thrombocytopenia) or by the Mpl^null mutation (a model of chronic thrombocytopenia) is also compared. The levels of TPO expressed by bone marrow and liver cells of the thrombocytopenic animals, as well as the levels of TPO protein present in plasma, is presented for comparison.

As expected [34,37], in mice, the major site of TPO mRNA production is the liver, and the expression of TPO in this organ is not affected by changes in the number of platelets in circulation (Table 1). The data presented in Table 1 also confirm that TPO mRNA is expressed in bone, kidney, and, to a lower extent, spleen of untreated wild-type mice. In addition, in these wild-type mice, TPO mRNA is found expressed by the intestine and stomach, although it is not detected in the ear. Mild bleeding significantly reduces the levels of TPO mRNA expressed by spleen, bone, and stomach but increases the levels of TPO mRNA expressed by the bone marrow and, to a surprise, by the ear. On the other hand, the presence of the Mpl^null mutation increases the levels of TPO mRNA levels expressed by the spleen and decreases those expressed by the bone, bone marrow, and intestine (Table 1).

Table 1.

TPO Expression in Tissues of Untreated and Bled Wild-Type Mice and of Mpl^nul1 Mice

Mice	Wild type	Bled wild type	Mplnull
Blood values	Plt (×l06)/μL:0.9(±0.03) Hct (%): 47.8(±0.3)	Plt (×l06)/μL: 0.7(±0.03)* Hct (%): 43.1(±1.1)*	Plt (×106)/μL: 0.03 (±0.004)* Hct (%): 42.3 (±1.3)*
	mRNA (×10−3 2−ΔCt)	mRNA (×10−3 2−ΔCt)	mRNA (×10−3 2−ΔCt)
Liver	39.3 ± 7	41.3 ± 4.3	35.5 ± 10.5
Spleen	1.3 ± 0.1	0.5 ± 0.2*	73.0 ± 18.0*
Bone marrow**	0.1 ± 0.006	3 ± 1*	0.01 ± 0.002*
Bone***	1.9 ± 0.5	0.5 ± 0.1*	b.d.
Ear	b.d.	1.2 ± 0.2*	b.d.
Kidney	6.2 ± 0.001	7.6 ± 0.9	8.9 ± 2.1
Stomach	0.7 ± 0.1	0.3 ± 0.03*	1.0 ± 0.1
Intestine	4.0 ± 0.2	8.2 ± 2.0	2.2 ± 0.1*

Results are presented as mean (±SD) of independent determinations with three mice per experimental group.

^*Results statistically different (P < 0.01) from those observed in untreated wild-type mice. b.d. = below detectable levels.

^**Bone marrow indicates mRNA prepared from cells harvested from the medullar cavity.

^***Bone indicates mRNA extracted from the frozen femurs grounded under liquid nitrogen vapors.

At protein levels, bleeding and the presence of the Mpl^null mutation significantly reduce the amount of TPO present in the blood (Table 2). This result is consistent with the observation that in humans the levels of TPO in the circulation are exquisitely sensitive to thrombocytopenia [36]. In addition, TPO protein is clearly detectable in washes from the marrow, liver, spleen, and peritoneal cavity (Table 2). Bleeding and the presence of the Mpl^null mutation increases and decreases, respectively, the levels of TPO protein present in spleen and liver washes.

Table 2.

Levels of TPO Protein Present in Plasma, in Organ Washes and in Peritoneal Lavages of Untreated and Bled Wild-Type Mice and of Mpl^nul1 Mice

	Wild-type mice**	Bled wild-type mice	Mplnul1mice
	Protein (pg/mL)	Protein (pg/mL)	Protein (pg/mL)
Plasma	1287 ± 205	451 ± 11*	351 ± 31*
Bone marrow washes	98 ± 13	96 ± 14	47 ± 13
Spleen washes	99 ± 15	140 ± 4*	119 ± 7
Liver washes	289 ± 28	179 ± 6*	271 ± 74
Peritoneum lavages	93 ± 2	88 ± 3	97 ± 15

^*Results statistically different (P < 0.01) from those observed in untreated wild-type mice.

^**The same mice analyzed in Table 1. Results are presented as mean (±SD) of independent determinations with three mice per experimental group.

These results indicate that TPO is expressed by a wide range of extramedullary sites, including those (intestine, gut, and skin) rich in mast cells and that the expression of TPO in these tissues is regulated by the number of platelets.

Discussion

We have suggested a model for the effects of TPO on mast cell differentiation in which TPO increases the number of mast cell precursors by favoring the expansion of the hematopoietic stem/progenitor cell compartment but decreases the total mast cell output by inducing apoptosis of the precursor cells that therefore cannot progress into maturation [27]. This model predicted that Mpl should be expressed on the surface of cells at all the stages of mastocytopoiesis and that targeted deletion of the Mpl gene should increase the number of mast cells found in the periphery. In this study, it is shown that this is indeed the case. In fact, robust levels of Mpl expression is detected using flow cytometry on the surface of MCP and mast cell precursors present in the marrow (Fig. 1). On the other hand, Mpl^null mice express decreased frequency of MCP and increased frequency of mast cell precursors in the marrow and spleen (Fig. 2). The Mpl^null phenotype also includes increased frequency of mast cells in the dermis, stomach, and peritoneal cavity (Figs. 3–8). Interruption of the Mpl–TPO interaction forced mast cell maturation as revealed by the more mature morphology of dermal (Fig. 4) and serosal (Fig. 8) mast cells from Mpl^null mice. In addition, serosal Mpl^null mast cells express high levels of MMCP-7 (Fig. 8), a protease characteristic of wild-type dermal mast cells [33]. This finding suggests that the Mpl^null mutation favors maturation and expansion of dermal-type mast cells and provides the first indication of the existence of factors, such as TPO, that may act as in vivo maturation switch between the different mast cell types.

The changes in the MMCP expression profile induced by the Mpl^null mutation are consistent with the current understanding of the molecular control of mastocytopoiesis: the high number of Mpl^null serosal mast cells reflecting that increased proliferation is associated with enhanced expression of Gata2, a gene that controls the proliferation of hematopoietic cells [18,38,39]. On the other hand, more mature features of the Mpl^null cells are reflected by increased expression of Gata1 and its target genes, MC-CPA [40] and FcɛRI [41,42], and of Mitf and its target genes, MMCP-6 [43] and Bcl2 [44] (Fig. 8).

Several positive regulators of mast cell differentiation have been identified. In addition to SCF and IL-3, IL-4 [45], IL-9 [46], and IL-16 [47], all play important positive regulatory functions of these lineages [48]. With the exception of glucocorticoids [49,50], however, negative regulators for mast cell differentiation have not been reported so far. Therefore, the overall effects of TPO on mast cell differentiation revealed by ectopic treatment of mice and mast cell cultures with this growth factor [27] and by targeted deletion of its receptor (this study) identify the first in vivo inhibitor of mastocytopoiesis.

The results provided in this article, and data obtained from the literature, allow speculation on the physiological role of TPO as a regulator of mast cell function. TPO is one of the most important in vivo regulators of megakaryocytic differentiation [15] as both Mpl^null [28] and TPO-treated [51] mice have altered levels of megakaryocytes in their marrow and of platelets in the blood. TPO is regulated by a negative feedback exerted by the platelets [25]. This negative loop ensures that changes in platelet concentrations due to exogenous stimuli, that is, bleeding, would result in increased TPO production in the marrow with resultant megakaryocyte production. An increase in the megakaryocyte population requires an increase in progenitor cell pool. However, the effects of TPO on the progenitor cell compartments are not selective as TPO increases the production of progenitor cells of all types [28,52] including the MCP (Fig. 2). Increase of these progenitor cells explain why, although modest, the numbers of all the hematopoietic precursors increase in the blood after TPO treatments [15]. Owing to the high reactivity of mast cells to exogenous stimuli, the number of mast cells must not be increased in this setting. This is ensured by the apoptotic-inducing effects of TPO on mast cell precursors [27]. Although allergic reactions do not involve bleeding, parasitic infections are associated with blood losses. It has been reported that experimentally induced infections with the worm Trichinella spiralis promote mast cell differentiation in the marrow [32]. However, the factors that increase the numbers of these precursors and that selectively amplify the serosal mast cells are unknown. As it is conceivable that this infection is also associated with TPO production in the marrow, the selective sensitivity of dermal (more sensitive) and serosal (less sensitive) mast cells to the apoptotic inducing effect of TPO might represent the physiological signal to selectively increase the number of serosal mast cells during parasite infections.

A corollarium of this hypothesis is that TPO should be expressed in extramedullary sites rich in mast cells, and that this expression should be sensitive to blood losses. It is currently accepted that TPO constitutively expressed by the endothelial cells of the liver represents the major source of growth factor in the plasma [34,37]. Megakaryocytes and platelets, by binding TPO and by internalizing and degrading the growth factor/receptor complex, play a major role in tuning the concentration of TPO in the plasma [53]. Accordingly, in humans, thrombocytopenias characterized by reduced megakaryocyte production (reduced Mpl pool) or increased platelet destruction (accelerated destruction of the TPO/Mpl complex) are associated, respectively, with increased and reduced TPO concentration in the plasma [36]. It is recently emerging that, in addition to liver, TPO expression is also detectable in bone marrow, spleen, and kidney by in situ hybridization [35]. An alternative spliced form of TPO is expressed in the human amygdala and hippocampus [54]. Although the functions exerted by TPO in spleen and kidney are not known, the expression of this growth factor in brain and bone marrow plays a major role in the regulation of neurogenesis and stem cell biology, respectively. In the brain, TPO induces apoptosis of newly generated neurons [55]. In the bone marrow, TPO is expressed by the osteoblasts of the endosteum [26,56] and by the stromal cells of the medulla [57]. TPO expression by the osteoblasts is supposed to contribute to the chemical architecture of the stem cell niche [26,56]. Quiescent stem cells associated with the osteoblasts are supposed to be induced to proliferate by the synergistic action of TPO and SCF, both produced by the osteoblasts and to leave the niche to complete their cycle. If stimulated with TPO again, the stem cell progeny enters into quiescence, returns to the osteoblast niche, and retain stemcellness [26,56]. In contrast to the constitutive TPO expression of the liver, expression of TPO in the marrow is exquisitely regulated by the platelets that establish a negative regulatory loop on the ability of the osteoblasts to produce TPO [25]. In addition, the granules within the platelets contain both factors that stimulate (platelet-derived growth factor [PDGF] and fibroblast growth factor 2 [FGF2]) and those that inhibit (platelet factor 4, thrombospondin, and TGF-β) TPO production by the stromal cells of the medulla [58]. The data presented in Tables 1 and 2 complement this information by indicating that TPO is produced by numerous extramedullary sites, including those (peritoneum, gut, and intestine) rich in mast cells. Platelet numbers not only regulates TPO concentration in plasma (both bled and Mpl^null mice express TPO blood levels lower than normal) but also exert a variegated effect on the levels of TPO expressed by cells of the bone, bone marrow, and extramedullary sites. In particular, both bleeding and the presence of the Mpl^null mutation reduce the levels of TPO expressed by bones, while bleeding increases the levels of TPO expressed by the medulla and to a surprise, those expressed in ears (Table 1).

In conclusion, these data confirm TPO as an important physiological negative regulator of mastocytopoiesis and identifies this growth factor as a possible effector of lineage switch between dermal and serosal mast cells.

Acknowledgments

This study was supported by a grant from the National Cancer Institute (P01-CA108671). Murine TPO and the Mpl-specific monoclonal AMM2 antibody were provided by Kirin (MTA of March 29, 2002; Gunma, Japan). Recombinant rat SCF was provided by Amgen (MTA No. 19982634–005; Thousand Oaks, CA). This work was partially presented at the Hematopoietic Stem Cell VI meeting (September 14–16, 2006), Tübingen, Germany.

Author Disclosure Statement

One of the authors (M.N.) was employed by a company (Kirin) whose potential products (Thrombopoietin and the AMM2 anti-Mpl antibody) were used in the present work. None of the other authors have potential conflict of interest.